A complete census of silicate features in the mid-infrared spectra of active galaxies

TL;DR

This study provides a comprehensive analysis of silicate features in nearly 800 active galactic nuclei, revealing how host galaxy subtraction affects feature detection and offering insights into dust distribution models in AGN.

Contribution

It presents the first large-scale, detailed measurement of silicate features in AGN spectra, emphasizing the importance of clumpy dust models and host galaxy subtraction.

Findings

Host galaxy subtraction increases detected emission features by 20-50%.

Silicate features show a bimodal peak distribution in emission cases.

Clumpy dust models better explain observed silicate features.

Abstract

We present a comprehensive study of the silicate features at 9.7 and 18 micron of a sample of almost 800 active galactic nuclei (AGN) with available spectra from the Spitzer InfraRed Spectrograph (IRS). We measure the strength of the silicate feature at 9.7 micron, S9.7, before and after subtracting the host galaxy emission from the IRS spectra. The numbers of type 1 and 2 AGN with the feature in emission increase by 20 and 50%, respectively, once the host galaxy is removed, while 35% of objects with the feature originally in absorption exhibit it in even deeper absorption. The peak of S9.7, lambda_peak, has a bimodal distribution when the feature is in emission, with about 65% of the cases showing lambda_peak > 10.2 micron. Silicates can appear in emission in objects with mid-infrared (MIR) luminosity spanning over six orders of magnitude. The derived distributions of the strength of the silicate features at 9.7 and 18 micron provide a solid test bed for modeling the dust distribution in AGN. Clumpiness is needed in order to produce absorption features in unobscured AGN and can also cause the silicates to be in absorption at 9.7 micron and in emission at 18 micron in type 1 sources. We find the `cosmic' silicates of Ossenkopf et al. to be more consistent with the observations than Draine's `astronomical' silicates. Finally, we discuss the possibility of a foreground absorber to explain the deep silicate absorption features in the MIR spectra of some type 2 AGN.